1896.] METALLIC COLOURS OF BIRDS. 293 
in the wing-quills in the direction of diminished efficiency for flight 
would be checked by natural selection. 
So far we have seen how the metallic colours both of Sun-birds 
and Humming-birds depend in each case upon a combination of a 
certain structure and a black pigment. In conclusion something 
may be said as to the colours themselves. In the Sun-birds a 
greenish-blue seems to be the most primitive metallic colour, and 
this is a very common tint elsewhere, e. g. Peacock, Quezal, &e. 
According to Gadow, one of the reasons why any metallic feather 
does not display all the colours of the spectrum is probably because 
the overlapping of successive colour-producing structures cuts out 
certain of the rays. If this overlapping really occurs it seems not 
unnatural to conclude that the middle rays of the spectrum, those 
in the neighbourhood of the green, would be least likely to be 
affected, and we would thus get green as a primitive metallic 
colour. The combination of this structure with a surface sculptur- 
ing might produce a purple or violet tint ; the absence of red and 
yellow may not improbably be a result of physical conditions. 
Walter explains the rarity of red and yellow metallic colours as 
due to the nature of the pigments contained in the coloured tissues, 
but this again is difficult to harmonize with our knowledge of such 
pigments. 
There can be little doubt that in Humming-birds a greenish- 
yellow is the most primitive metallic tint. It is suggested even in 
the “hermit” forms, and is very widely spread on contour-feathers 
elsewhere. In the absence, however, of any suggested physical 
explanation of the metallic colours of Humming-birds, it would 
perhaps be premature to attempt to account for the wonderful 
range of colour found in the family. As to the distribution of 
metallic colour one or two facts still remain to be noticed. hus 
metallic colour is not always characteristic of the male. In Huste- 
phanus fernandensis female metallic colour is more or less distinctly 
present over nearly the whole of the upper and under surfaces, 
the tail-quills show bright metallic colour, and the head bears a 
special metallic crest. The male, on the other hand, has no metallic 
colour except the bright crest, the rest of the body is cinnamon- 
coloured and without metallic gloss. The absence of metallic 
colour is apparently to be accounted for here by the absence of the 
usual blackish-brown pigment. Again, a specimen marked Topaza 
pella, young male, which was examined, showed metallic-greenish 
feathers in the upper part of the head, a spot which in the adult 
is covered with black feathers. Similarly, Salvin notices that in 
Lampornis mango the throat in the young bird is covered with 
glittering green feathers, and in the adult with pure black ones. 
Thus apparently an excess of black pigment is as fatal to the 
display of metallic colour as its total absence. As to the relation 
between a black colour and metallic tints there are some other 
interesting facts. In Oyanolesbia gorgo the tail-quills are greatly 
elongated and show gorgeous metallic colour, but this is confined 
to the distal end of the feathers, the proximal region being a 
